Early Posting

Accepted papers to appear in an upcoming issue

Optica Publishing Group posts prepublication articles as soon as they are accepted and cleared for production. See the FAQ for additional information.

No-Reference Image Quality Metrics for ColorDomain Modified Images

Muhammad Usman Khan, Ming Luo, and Dalin Tian

DOI: 10.1364/JOSAA.450595 Received 07 Dec 2021; Accepted 16 May 2022; Posted 16 May 2022  View: PDF

Abstract: Predicting the quality of natural images without using a reference image has always been a challenging task. Numerous approacheshave been proposed in the past, but they mainly focused on spatial and frequency domain degradations like blur, noise,compression, etc. The Image Quality Metrics (IQMs) in literature perform with quite a high accuracy for such types of degradedimages. However, their performances were not good on the images modified in the color domain. In this study, psychophysicalexperiments were conducted to assess the image quality of the color domain images. Two datasets were generated here andtogether with a previously published dataset were used to develop three IQMs based on absolute values, relative values, andstatistical analysis of image color appearance attributes. Their performances were then evaluated together with five spatialdomain IQMs from the literature. The results showed that the color-domain IQMs outperformed the others. The absolute andrelative attributes-based models, when combined, achieved the best performance. The present results suggest that more effort isneeded to improve the performance of color domain IQMs for image quality estimation.

Degree of paraxiality of a twistelectromagnetic Gaussian Schell-model beam

Ju Huang, Wen Jiang, jinping Cheng, Xiaoling Ji, and Tao Wang

DOI: 10.1364/JOSAA.456685 Received 21 Feb 2022; Accepted 16 May 2022; Posted 16 May 2022  View: PDF

Abstract: The definition of the degree of paraxiality (DOP) for stochastic electromagneticfield is applied to twist stochastic electromagnetic field. As an illustrative example, DOP for awide class of model stochastic field, i.e., twist electromagnetic Gaussian Schell-model(TEGSM) field, is discussed. The dependence of DOP of the light source on its properties hasalso been studied in detail. The numerical results have shown that DOP of a TEGSM beam isdetermined by the rms widths of the auto-correlation functions and the twist factor as well asby the degree of polarization. In addition, in order to explain the behavior of DOP, the farfield divergence angle of this beam source is also discussed.

Bounds on mutual information of mixture data forclassification tasks

Yijun Ding and Amit Ashok

DOI: 10.1364/JOSAA.456861 Received 22 Feb 2022; Accepted 12 May 2022; Posted 13 May 2022  View: PDF

Abstract: To quantify the optimum performance for classification tasks, the Shannon mutual information is a naturalinformation-theoretic metric, as it is directly related to the probability of error. The data produced by manyimaging systems can be modeled by mixture distributions. The mutual information between mixture dataand the class label does not have an analytical expression, nor any efficient computational algorithms.We introduce a variational upper bound, a lower bound, and three approximations, all employing pairwise divergences between mixture components. We compare the new bounds and approximations withMonte Carlo stochastic sampling and bounds derived from entropy bounds. To conclude, we evaluate theperformance of the bounds and approximations through numerical simulations.

Is Melanopsin Activation Affecting Large Field Color Matching Functions?

Pablo Barrionuevo, Clemente Paz Filgueira, and Dingcai Cao

DOI: 10.1364/JOSAA.457223 Received 28 Feb 2022; Accepted 11 May 2022; Posted 11 May 2022  View: PDF

Abstract: Color theory is based on the exclusive activation of cones. However, since the discovery of melanopsin expressing cells in the human retina, evidence of its intrusion in brightness and color vision is increasing. We aimed to assess if differences between peripheral or large field and foveal color matches can be accounted for melanopsin activation or rod intrusion. Photopic color matches by young observers showed that differences between extrafoveal and foveal results cannot be explained by rod intrusion. Furthermore, statistical analyses on existing color matching functions suggest a role of melanopsin activation, particularly, in Large Field S Fundamentals.

UMA-Net: an unsupervised representation learningnetwork for 3D point cloud classification

jie liu, yu tian, GUOHUA Geng, haolin wang, da song, kang li, mingquan zhou, and Xin Cao

DOI: 10.1364/JOSAA.456153 Received 14 Feb 2022; Accepted 07 May 2022; Posted 11 May 2022  View: PDF

Abstract: The success of deep neural networks usually relies on massive amounts of manually labeled data, which is bothexpensive and difficult to obtain in many real-world datasets. In this paper, a novel unsupervised representationlearning network, UMA-Net, is proposed for the downstream 3D object classification. Firstly, the multi-scale shell-basedencoder is proposed, which is able to extract the local features from different scales in a simple yet effective manner.Secondly, an improved angular loss is presented to get a good metric for measuring the similarity between local featuresand global representations. Subsequently, the self-reconstruction loss is introduced to ensure the globalrepresentations do not deviate from the input data. Additionally, the output point clouds are generated by the proposedcross-dim-based decoder. Finally, a linear classifier is trained using the global representations obtained from the pretrained model. Furthermore, the performance of this model is evaluated on ModelNet40 and applied to the real-world3D Terracotta Warriors fragments dataset. Experimental results demonstrate that our model achieves comparableperformance and narrows the gap between unsupervised and supervised learning approaches in downstream objectclassification tasks. Moreover, it is the first attempt to apply the unsupervised representation learning for 3D TerracottaWarriors fragments. We hope this success can provide a new avenue for the virtual protection of cultural relics.

Towards non-metameric reflectance recovery byemulating spectral neighborhood using correspondingcolor information

Muhammad Safdar and Patrick Emmel

DOI: 10.1364/JOSAA.451931 Received 22 Dec 2021; Accepted 06 May 2022; Posted 06 May 2022  View: PDF

Abstract: In learning-based reflectance reconstruction methods, usually localized training samples are used to reconstruct spectral curves. The state-of-the-art methods localize the training samples based on their colorimetric color differences with the test sample. This approach is dependent on the working color space,color difference equation, and/or illuminant used, and may result in a metameric match. This issue canbe resolved by localizing the training samples based on their spectral difference with the test sample,however, this would require an already unknown spectral curve of the test sample. In this paper, use ofcorresponding color information to emulate spectral neighborhood of the test color for non-metamericreflectance recovery is proposed. The Wiener estimation method was extended by (1) using two thresholds; (i) on the color difference between the test sample and the training samples under the referenceilluminant, and (ii) on the color difference between the corresponding color of the test sample and thetraining samples under another illuminant, to mimic the spectral neighborhood of the test sample withinthe gamut of the training data, and (2) also using the tristimulus values of the corresponding color in theregression. Results showed that the proposed extension of the Wiener estimation method improved thereflectance recovery and hence reduced the metamerism.

Autoencoder Based Training for Multi Illuminant Color Constancy

Donik Vrsnak, Ilija Domislovic, Marko Subašić, and Sven Lončarić

DOI: 10.1364/JOSAA.457751 Received 07 Mar 2022; Accepted 06 May 2022; Posted 10 May 2022  View: PDF

Abstract: Color constancy is an essential component of the human visual system. It enablesus to discern the color of objects invariant to the illumination that is present. This ability isdifficult to reproduce in software as the underlying problem is ill-posed, i.e. for each pixel in theimage we only know the RGB values which are a product of the spectral characteristics of theillumination and the reflectance of objects, as well as the sensitivity of the sensor. In order tocombat this, additional assumptions about the scene have to be made. These assumptions caneither be handcrafted or learned using some deep learning technique. Nonetheless, they mostlywork only for single illuminant images. In this work, we propose a novel method for learningthese assumptions for multi illuminant scenes using an autoencoder trained to reconstruct theoriginal image by splitting it into its illumination and reflectance components. We then showthat the estimation can be used as is or can be used alongside a clustering method to createa segmentation map of illuminations. We show that our method performs the best out of alltested methods in multi illuminant scenes while being completely invariant to the number ofilluminants.

The Point Spread Function of the Polarized Light Field Microscope

Mai Tran and Rudolf Oldenbourg

DOI: 10.1364/JOSAA.458034 Received 10 Mar 2022; Accepted 03 May 2022; Posted 09 May 2022  View: PDF

Abstract: We examined the point spread function of the polarized light field microscope andestablished a computational framework to solve the forward problem in polarized light fieldimaging, for the purpose of furthering its use as a quantitative tool for measuring 3-dimensionalmaps of the birefringence of transparent objects. We recorded experimental polarized light fieldimages of small calcite crystals and of larger birefringent objects and compared ourxperimental results to numerical simulations based on polarized light ray tracing. We findgood agreement between all our experiments and simulations, which leads us to proposepolarized light ray tracing as one solution to the forward problem for the complex, non-linearimaging mode of the polarized light field microscope. Solutions to the ill-posed inverseproblem might be found in analytical methods and/or deep learning approaches that are basedon training data generated by the forward solution presented here.

Sharp-edge diffraction under Bessel beam illumination. A catastrophe optics perspective

Riccardo Borghi and Cristian Carosella

DOI: 10.1364/JOSAA.455654 Received 09 Feb 2022; Accepted 02 May 2022; Posted 06 May 2022  View: PDF

Abstract: Boundary Diffraction Wave theory and Catastrophe Optics have already proved to bea formidable combination in computational optics. In the present paper a general paraxial theory aimed at dealing paraxial diffraction of Bessel beams by arbitrarily shaped sharp-edge apertures is developed. A key ingredient of our analysis is the 𝛿-like nature of the angular spectrum of nondiffracting beams. This allows the diffracted wavefield to be effectively represented through two-dimensional integrals defined onto rectangular domains, whose numerical evaluation iseasily achievable via standard Montecarlo techniques. As a byproduct of the present analysis, asimple explanation of a recently observed property of some “heart-like” apertures to flatten the axial intensity of apodized Bessel beams is also provided.

Orbital angular momentum of paraxial propagation-invariant laser beams

Victor Kotlyar and Alexey Kovalev

DOI: 10.1364/JOSAA.457660 Received 04 Mar 2022; Accepted 28 Apr 2022; Posted 05 May 2022  View: PDF

Abstract: For propagation-invariant laser beams represented as a finite superposition of theHermite-Gaussian beams with the same Gouy phase and with arbitrary weight coefficients,we obtain an analytical expression for the normalized orbital angular momentum (OAM).This expression is represented also as a finite sum of weight coefficients. We show thatcertain choice of the weight coefficients allows obtaining the maximal OAM, which is equalto the maximal power of the Hermite polynomial in the sum. In this case, the superpositiondescribes a single-ringed Laguerre-Gaussian beam with a topological charge equal to themaximal OAM and to the maximal power of the Hermite polynomial.

Into the wild of non-linear electromagnetism: tutorial.A course on non-linear electromagnetism, not quitefrom scratch– Part I

Frederic Zolla and Pierre Godard

DOI: 10.1364/JOSAA.442439 Received 10 Sep 2021; Accepted 28 Apr 2022; Posted 13 May 2022  View: PDF

Abstract: This tutorial is aimed at introducing in a natural fashion the propagation equations system governingthe scattering of an electromagnetic wave by a nonlinear medium. The purpose is first to obtain the equations showing the most common nonlinear effects such as Kerr effect or the second and the third harmonicgeneration by avoiding conventional recipes and trying to arrive at these equations with a minimum ofassumptions. For this, we start from the general Maxwell’s equations involving the fields E, B, D andH and we rigorously provide all the hypotheses needed to attain the nonlinear systems of PDEs involving the different complex amplitudes of the different fields associated with the different frequencies atstake. In part II, the difficult question of energy transfer between fields emitted at the various frequenciesinvolved is discussed in detail. We then examine the tensorial nature of susceptibilities and, using Neumann’s principle, show how the number of their independent components can be significantly reduced.In part III, numerical examples of scattering by non-linear materials are given and discussed.

Into the wild of non-linear electromagnetism: tutorial. Acourse on non-linear electromagnetism, not quite fromscratch– Part II

Frederic Zolla

DOI: 10.1364/JOSAA.442739 Received 10 Sep 2021; Accepted 28 Apr 2022; Posted 13 May 2022  View: PDF

Abstract: This tutorial is aimed at introducing in a natural fashion the propagation equations system governing thescattering of an electromagnetic wave by a nonlinear medium. The purpose is first to obtain the equationsshowing the most common nonlinear effects such as Kerr effect or the second and the third harmonicgeneration by avoiding conventional recipes and trying to arrive at these equations with a minimum ofassumptions. For this, we start from the general Maxwell’s equations involving the fields E, B, D and Hand we rigorously provide all the hypotheses needed to attain the nonlinear systems of PDEs involving thedifferent complex amplitudes of the different fields associated with the different frequencies at stake. Inpart II, the difficult question of energy transfer between fields emitted at the various frequencies involvedis discussed in detail. We then examine the tensorial nature of susceptibilities and, using Neumann’sprinciple, show how the number of their independent components can be significantly reduced. In part III,numerical examples of scattering by non-linear materials are given and discussed.

Into the wild of non-linear electromagnetism : tutorial.A course on non-linear electromagnetism from“almost” scratch – Part III

Frederic Zolla and Pierre Godard

DOI: 10.1364/JOSAA.442743 Received 10 Sep 2021; Accepted 28 Apr 2022; Posted 13 May 2022  View: PDF

Abstract: This tutorial is aimed at introducing in a natural fashion the propagation equations system governing the scattering of an electromagnetic wave by a nonlinear medium. The purpose is to let the effects appear spontaneously deleting as far as possible the \textit{ad hoc} recipes. In this way, propagation equations systems that encode the most of nonlinear effects are obtained. This aim is partly reached especially when treating harmonic generations. For this, we start from the general Maxwell's equations involving the fields $\mathbf{E}$, $\mathbf{B}$, $\mathbf{D}$ and $\mathbf{H}$ and we give rigorously all the hypotheses needed to attain the nonlinear systems of PDEs involving the different complex amplitudes of the different fields associated with the different frequencies at stake. In addition, in the second part, the difficult question of energy transfers between the different fields emitted at the different frequencies involved is discussed in detail. We then examine the tensorial nature of susceptibilities and, using Neumann's principle, discover how the number of their independent components can be significantly reduced. In the third part, numerical examples of scattering by non-linear materials are given and discussed.

Performing tomographic reconstructions froma satellite looking toward Earth. Part 1:implementation and limitations

Garrett Hinton, Harri Latvakoski, Mike Taylor, Eric Clarkson, Matthew Kupinski, and Jed Hancock

DOI: 10.1364/JOSAA.449217 Received 23 Nov 2021; Accepted 25 Apr 2022; Posted 02 May 2022  View: PDF

Abstract: For imaging instruments that are in space looking toward the Earth, there are avariety of nuisance signals that can get in the way of performing certain imaging tasks, such asreflections from clouds, reflections from the ground, and emissions from the OH-airglow layer.A method for separating these signals is to perform tomographic reconstructions from thecollected data. A lingering struggle for this method is altitude-axis resolution and differentmethods for helping with it are discussed. An implementation of the Maximum LikelihoodExpectation Maximization (MLEM) algorithm is given and analyzed.